From Cancer Biology to Drug Interactions: Machine Learning for Complex Biomedical Problems

Progress in biomedicine increasingly depends on our ability to interpret complex and heterogeneous data. In cancer, disease development is shaped by interactions among genetic alterations, molecular pathways, and clinical variables. Similarly, combinations of drugs may lead to harmful interactions that are difficult to anticipate using conventional approaches. These challenges call for computational methods that can learn from structured, high-dimensional data.

In this docent lecture, I will discuss how machine learning can be used to address these problems, with examples from cancer research and drug interaction prediction. Particular emphasis will be placed on graph-based and multimodal models, and on how these approaches can support biomarker discovery, improve our understanding of disease mechanisms, and predict harmful drug combinations. The lecture highlights how machine learning can contribute to a more systematic understanding of complex biomedical problems and support progress in precision medicine and translational research.